Merchandising rules

Hand-pick which products to push, hide, or downrank — even when the AI thinks otherwise. Your store, your call.

CartAmplify’s AI usually does a great job ordering products for each shopper. But sometimes you know better than the AI: a new arrival you want everyone to see, a slow-moving product you want pushed down, a discontinued item you need hidden until the new version lands. Merchandising rules give you these manual overrides without disabling personalization for everyone else.

The three rule types#

Boost

Push specific products higher in search, browse, or recommendation results. Use for new arrivals, exclusives, items you want to clear, paid placements.

Bury

Push specific products lower without hiding them. Use for slow-movers you don’t want to deprioritize entirely, items with low margin, last-season inventory.

Hide

Remove specific products from results entirely. Use for region-restricted items, products awaiting compliance approval, or temporary holds.

Where each rule applies#

When you create a rule, you choose which surfaces it affects:

SurfaceUse case
Search resultsBoost a “best seller” for every relevant query, hide a discontinued line
Browse / category pagesPush your “Summer Sale” hero items to the top of the Sale category
Recommendation widgetsBury low-margin SKUs in the cart-page cross-sell, boost a new product in “Recommended for you”

Recommendation rules can be scoped further — apply to all recommendation widgets, or only specific ones (e.g., only the homepage “Recommended for you” shelf, not the PDP “Similar items”).

Targeting & scheduling#

Each rule has a few knobs that decide when and where it fires:

  • Category filter — apply the rule to all categories, or only to specific ones. Useful when “boost this product” should only happen on relevant pages.
  • Schedule — always-on, or a date range. A “Black Friday” boost rule can be set to activate at midnight on the day and automatically expire when the sale ends.
  • Language — like the rest of CartAmplify, rules are scoped per language. A boost rule on your English store doesn’t affect your Spanish store.
  • Strength (for boost / bury) — a slider that controls how aggressive the override is. A gentle boost nudges a product up a few positions; a strong boost pushes it near the top.

How rules interact with the AI#

This is the most important thing to understand. Merchandising rules don’t replace the AI — they adjust its output.

Rules apply after the personalization model has ranked products, and they apply proportionally. Concretely:

  • Boost multiplies the product’s relevance score, but the AI’s underlying signal still matters — a heavily-boosted low-relevance product can’t always leapfrog a strong organic match. This protects shoppers from seeing wildly irrelevant items just because you boosted them.
  • Bury does the opposite — multiplies down the relevance.
  • Hide is the only absolute override: hidden products never appear, regardless of relevance.

The intent: you can tilt the playing field, but you can’t break it. This is how you avoid the classic merchandising trap where “everyone boosts everything” and personalization stops working.

Setting up a rule#

In Dashboard → AI Features → Merchandising AI → Rules → New rule, the workflow is:

  1. Pick a rule type — boost, bury, or hide.
  2. Pick the products — by SKU, by category, by brand, by tag. You can pin a single product or apply to a whole collection.
  3. Pick the surfaces — search, browse, recommendations (with per-widget toggles for recommendations).
  4. Set the scope — all categories or specific ones; for boost/bury, the strength slider.
  5. Schedule it — always-on, or pick start and end dates.
  6. Activate — the rule starts affecting results within minutes.

You can pause or delete a rule any time. Pausing keeps the rule in place but stops applying it — useful for seasonal rules you’ll want back next year.

Watching your rules in action#

Once rules are live, Dashboard → AI Features → Merchandising AI → Analytics shows:

  • Per-rule impressions, clicks, and revenue impact
  • Which rules are firing on which queries / pages
  • Comparison vs the AI-only baseline — did boosting this product actually lift conversion, or did it cannibalize a stronger organic ranker?

This is the feedback loop. A boost rule that drives revenue stays on; one that hurts conversion should be paused or relaxed.

Common patterns#

Campaign-driven boosts

“Push the new fall collection to the top of the Women’s category for the next two weeks.” Boost rule, category filter = Women’s, scheduled with start/end dates.

Inventory clearance

“Move these 12 SKUs out by end of quarter.” Boost rule pinned to those SKUs across search and browse, expires at quarter end.

Brand priority

“Always show our own-label products before competitor brands on category pages.” Boost rule scoped to brand=YourLabel, applied to browse only.

Compliance hide

“Hide product SKU-XYZ in the EU until it clears review.” Hide rule pinned to that SKU, scoped to the EU-language storefronts.

Common questions#

Where to go next#